Introduction to Quantitative Research Methods_ An Investigative Approach.pdf

1. INTRODUCTION TO QUANTITATIVE RESEARCH METHODS

2. INTRODUCTION TO QUANTITATIVE RESEARCH METHODS A N I N V E S T I G A T I V E APPROACH M A R K B A L N A V E S a n d P E T E R C A P U T I SAGE Publications London . Thousand Oaks . New Delhi

3. # Mark Balnaves and Peter Caputi 2001 Introduction # Alec McHoul 2001 First published 2001 Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted in any form, or by any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency. Inquiries concerning reproduction outside those terms should be sent to the publishers. SAGE Publications Ltd 6 Bonhill Street London EC2A 4PU SAGE Publications Inc 2455 Teller Road Thousand Oaks, California 91320 SAGE Publications India Pvt Ltd 32, M-Block Market Greater Kailash-I New Delhi 110 048 British Library Cataloguing in Publication data A catalogue record for this book is available from the British Library ISBN 0-7619-6803-2 ISBN 0-7619-6804-0 (pbk) Library of Congress catalog record available Typeset by Keyword Publishing Services Limited, UK Printed in Great Britain by The Cromwell Press Ltd, Trowbridge, Wiltshire

4. Contents Acknowledgements xiii Permissions xiv 1 Order at all Points: Counting and accounting 1 The Ql-Qt Continuum 1 Culture and the Reliability of `Data' 5 References 8 2 Starting the Inquiry: `But what happened then?' 10 Knowing What to Research 16 Exploration, Description and Explanation 16 Topics of Research 20 Knowing Whom to Research 21 Units of Analysis 21 Knowing How to Research 22 Literature Review 22 Secondary Analysis 25 Knowing When to Research 25 The Importance of Time 25 Research design 27 Summary 27 The Statistical Inquirer 29 Main Points 29 Review Exercises 30 References 31 3 De®ning the Inquiry: `Then how do you know?' 33 Tools of Methodology 34 Styles of Reasoning (deduction, induction and abduction) 36 Causality 40 Mapping Relationships 41 Tools of Measurement 44 What are Variables? 45 Operational Definitions 48 Contemporary Views on Measurement 50 Great Sociological Detective Stories: Collecting Data Across Cultures: Can we measure cultural variation? Culture's Consequences (Geert Hofstede) 51 Methodology and Theory 52 Method 53 Hypotheses and Operationalization 53 v

5. C O N T E N T S Variables 55 Individualism and Masculinity 55 Power Distance 56 Uncertainty Avoidance 57 Summary 59 Main Points 61 Review Exercises 61 References 62 4 Methods of Inquiry: `It is a capital mistake to theorize before one has data!' 64 Involvement and Method 65 Experimental Design 67 Different Types of Experimental Design 74 Survey Design 75 What is a Survey? 76 The Variable in Question 77 Nominal Variables/Questions 77 Ordinal Variables/Questions 78 Interval Variables/Questions 78 Ratio Variables/Questions 79 Multiple-Item Scales 79 Differential Scales 79 Cumulative Scales 80 Summative Scales 80 The Words in the Question 81 Administering the Questionnaire 84 Layout 84 Length of Questionnaire 85 Interview as Measurement 86 Validity 87 External Validity and Sampling 90 Population and Sampling Frame 91 Simple Random Sampling 91 Stratified Random Sampling 92 Multi-Stage Cluster Sampling 92 Sample Size 93 Non-Probability Sampling 95 Trade-Offs 95 Triangulation 95 Great Media and Politics Detective Stories: Using Survey Data. Do media change people's political attitudes? The People's Choice (Paul Lazarsfeld) 96 Methodology and Theory 97 Method 98 Hypotheses and Operationalization 98 Variables 100 vi

6. C O N T E N T S Summary 103 Main Points 105 Review Exercises 106 References 106 5 `Data! Data! Data!': Analysing data from the inquiry 109 Looking at the Clues: The Statistical Sleuth 109 Why Explore Data? 110 Plotting Data 110 Stem and Leaf Displays 110 Example 5.1: Stem and leaf display 111 Histograms 113 Boxplots 115 Tables, Graphs and Figures 118 Does a Picture Always Paint a Thousand Words? Some issues with representing data in graphical and tabular form 120 Using SPSS and Excel to Plot Data: Accounting for Tastes dataset 122 Methodology and Operationalization 123 Working with SPSS 124 Working with Excel 131 Describing Data 132 Numerical Summaries: What are they and why are they important? 132 Describing Location or Centrality 133 Some Notation 133 Example 5.2: Calculating the median of odd and even numbered distributions 134 Mean vs Median (When is the evidence not contaminated?) 134 Describing Variability 135 Example 5.3: Calculating the variance of a set of values 137 Using SPSS and Excel to Describe Data 138 Working with SPSS 138 Working with Excel 140 Summary 142 Main Points 143 Review Exercises 144 References 145 6 Finding Answers from the Inquiry: `Elementary, my dear Watson!' 147 Looking at Bivariate Data, Correlation and Regression 149 Plotting Bivariate Data 149 Correlation: a Measure of Co-Relation 151 Example 6.1: Computing the correlation coefficient 154 Introduction to Simple Linear Regression 156 Prediction and Correlation 156 Method of Least Squares 157 Example 6.2: Finding the regression line 159 Assessing the Fit of the Regression Model 160 Issue of Causation 161 vii

7. Using SPSS: Correlation and Regression 163 Using Excel 168 Looking at Categorical Data 171 Exploring Bivariate Categorical Data 171 Example 6.3: Computing the contingency coefficient 174 Inference: From Samples to Populations 175 Parameters, Estimates and Statistics 176 Sampling Distributions 177 Hypothesis Testing ± Don't just show me the evidence, convince me that it is so! 180 One-Sample Tests 180 One-Sample Test for Categorical Data 183 Two-Sample Tests 185 Test of Equality of Means for Independent Samples 185 Example 6.4: t-test for independent samples 186 Rank Test for Independent Samples 189 Example 6.5: Rank test for two independent groups 190 Tests for Categorical Data 191 Tests for Related Samples 193 Rank Test for Correlated or Related Samples 195 Using SPSS and Excel: One-Sample Tests 196 Using SPSS and Excel: Two-Sample Tests 201 Introduction to Randomization Tests 211 Great Social Biology Detective Stories: `We must hunt for cases' 213 Is There An Ideal Person?: Quetelet, Galton, Pearson 214 Does a General Intelligence Factor Exist?: Spearman and the introduction of correlation 215 Is There a Genetic Component to Intelligence? Did Burt clone the data? ± data ain't always data!! 217 Are Social Factors More Important than Individual Inclinations? Durkheim's study on suicide (The use of secondary data) 218 Summary 223 Main Points 224 Review Exercises 225 References 227 7 Summarizing and Presenting Results: `Quite so!' 229 Reporting Empirical Research 232 Writing Style, Narrative Style 234 Detection and Deception 235 References 236 Appendix I: Sample Letter for Informed Consent 237 Appendix II: BSA Statement of Ethical Practice 239 Appendix III: The Statistical Inquirer 246 Index 249 C O N T E N T S viii

8. Tables Table 2.1 Impact of television on Western Samoan evening activities, 4±10 pm 20 Table 2.2 Time and research design 26 Table 2.3 Organizing a quantitative research study 28 Table 3.1 From construct to operational definition 54 Table 3.2 Actual questions used to construct individualism/masculinity indexes 54 Table 3.3 Countries ranked by individualism scores 56 Table 3.4 Countries ranked by masculinity scores 56 Table 3.5 Countries ranked by power distance scores 57 Table 3.6 Countries ranked by uncertainty avoidance scores 58 Table 4.1 Effectiveness of Program A and B for men with initially unfavourable and men with initially favourable attitudes 74 Table 4.2 Effectiveness of Program A and B for men of different educational backgrounds 74 Table 4.3 Examples from Thurstone's differential scale 80 Table 4.4 Examples from Bogardus Social Distance scale 80 Table 4.5 Selected examples from Christie and Geis's Likert scale 81 Table 4.6 Sample questionnaire and coding column 86 Table 4.7 Rank order of French high school students' intelligence scores obtained with three hypothetical measures 89 Table 4.8 Table of random numbers 92 Table 4.9 Random selection of households for interview 94 Table 4.10 Summary of sampling procedures 94 Table 5.1 Example of back-to-back plot 113 Table 5.2 Frequency distribution table for grouped data 114 Table 5.3 Hypothetical data for variables X and Y 117 Table 5.4 The anatomy of a table 120 Table 5.5 Mean ratings of intensity of emotion 122 Table 5.6 Reframed data: mean ratings of intensity of emotion 122 Table 5.7 Accounting for tastes: comparison of stratified sample with official statistics 124 Table 6.1 Hypothetical data 149 Table 6.2 Hypothetical data on correlation between years of counselling experience and effective outcome 154 Table 6.3 Data with outliers 156 Table 6.4 Distribution of votes by section 172

9. Table 6.5 Percentage of votes within each section 172 Table 6.6 A 2 2 contingency table 173 Table 6.7 Data for job performance and EC test 174 Table 6.8 Computational details for Table 6.7 175 Table 6.9 Some commonly used values from a set of normal tables 182 Table 6.10 Frequencies in colour and shape of peas 184 Table 6.11 Hypothetical data for number of successful free throws in two conditions 185 Table 6.12 Descriptive statistics 187 Table 6.13 Frequencies relating to marital status and IQ 192 Table 6.14 Comparison of observed and expected frequencies for Baller's data 192 Table 6.15 Hypothetical data for speech anxiety scores in two conditions 194 Table 6.16 All possible permutations of six scores in two conditions 212 Table 6.17 Permutation distribution for data in Table 6.16 213 Table 6.18 Suicides in different countries per million persons of each confession 220 Table 6.19 Influences of the family on suicide, unmarried men 221 Table 6.20 Influences of the family on suicide, unmarried women 221 Table 7.1 Example from Accounting for Tastes: combined music genre preferences by gender 233 TA B L E S x

10. Figures Chart 1.1 Reports on Aboriginal youth and all youth-crime reports, Feb. 1991 to Jan. 1992 3 Chart 1.2 News reports on Aboriginal youth-crime and all youth-crime with actual crime data, Feb. 1991 to Jan. 1992 4 Figure 2.1 From assertions to evidence 23 Figure 3.1 Operationalization 48 Figure 4.1 Methods of data collection and personal involvement 65 Figure 4.2 Checklist for research design 66 Figure 4.3 Research methods and techniques of data collection 67 Figure 4.4 Basic experimental design 69 Figure 4.5 Different answering formats 84 Figure 4.6 Contingency questions 84 Figure 4.7 Factors affecting people's motivation to provide complete and accurate information to the interviewer 87 Figure 4.8 Checklist for questionnaire design 88 Figure 4.9 Multi-stage cluster sampling ± following the census tracts 93 Figure 4.10 Experimental choice based on issues of internal and external validity 96 Figure 4.11 Whereas actual occupation does little to refine the relationship between SES level and vote, it makes more difference whether a voter considers himself as belonging to `business' or `labour' 100 Figure 4.12 Religious affiliation splits the vote sharply 101 Figure 4.13 One-step model of mass-media influence 102 Figure 4.14 Two-step model of mass-media influence 102 Figure 4.15 Inductive approach 104 Figure 4.16 Deductive approach 104 Figure 5.1 A stem and leaf display of ESP data 114 Figure 5.2 Histogram of hypothetical examination marks 115 Figure 5.3 The anatomy of a boxplot 116 Figure 5.4 Boxplots for two hypothetical variables X and Y 117 Figure 5.5 Side-by-side boxplots 118 Figure 5.6 Preference for telecommunications carrier 121 Figure 5.7 A different way to display preference for telecommunications carrier 121 Figure 6.1 Scatterplot for data in Table 6.1 150

11. Figure 6.2 Negative association between two variables 150 Figure 6.3 No relationship between two variables 151 Figure 6.4 Scatterplot showing an outlier 151 Figure 6.5 A curvilinear relationship 155 Figure 6.6 Fitting a line through points on a scatterplot 157 Figure 6.7 Illustrating the concept of residual 158 Figure 6.8 Bar chart of frequency of votes by selection 172 Figure 6.9 Bar chart of raw percentage of votes by selection 173 Figure 6.10 The normal distribution of a normal curve 179 Figure 6.11 The t-distribution 183 Figure 6.12 Durkheim's theoretical hypothesis on suicide 220 Figure 6.13 Durkheim's hypothesis of degree of integration 221 Figure 6.14 Durkheim's hypothesis of degree of regulation 222 F I G U R E S xii

12. Acknowledgements A special thanks to: Gary Bouma and David DeVaus Mark Busani, Nick Castle and Monica Vecchiotti for their help with the multimedia courseware. Maurice Dunlevy, for contributions on journalism, and Harry Oxley, for contributions on causal diagramming. Patrick Rawstorne for use of his PhD dataset Predicting and Explaining the use of Information Technology with Value Expectancy Models of Behaviour in Contexts of Mandatory Use. Erika Pearson ± it's hard to find good help nowadays. Tony Bennett, Mike Emmison and John Frow for use of their dataset from The Australian Everyday Consumption project. SPSS illustrations have been reprinted by SPSS copyright permission. Excel illustrations have been reprinted by Excel copyright permission. The Apple University Consortium and PCTech for their equipment and software support. The Australian National Library for assistance with access to The Strand, from which the original Paget sketches of Sherlock Holmes were reproduced for this book. Alec McHoul, Mike Innes, Joyce and Michele Balnaves, Wendy Parkins and James Donald, who provided valuable insights into detection. John and Paul Balnaves on questions of Shakespeare and logic. Michele, Mary-Claire, Gerard, Elayne, James, and Jack. Douglas Adams ± the bottle of red has been sent. The authors of detective fiction.

13. Permissions The authors and publishers wish to thank the following for their permission to use copyright material: Tables 6.16 and 6.17. Copyright # 1993. Canadian Psychological Association. Reprinted with permission. Page 12, extract from The Strange Crime of John Boulnois by G.K. Chesterton. Used by permission of A.P. Watt on behalf of The Royal Literary Fund. Page 14, extract from Dirk Gently's Holistic Detective Agency by Douglas Adams. Used by permission of Douglas Adams. Copyright # 1987. Heinemann. Figure 2.1. From Evaluating Social Science Research, second edition, by Paul C. Stern and Linda Kalof. Copyright # 1979, 1996. Oxford University Press, Inc. Used by permission of Oxford University Press. Table 2.2. Used by permission of Gary Bouma. Copyright # 1993. The Research Process. Oxford University Press. Page 38, extract from The Blue Cross by G.K. Chesterton. Used by permission of A.P. Watt on behalf of The Royal Literary Fund. Figure 4.3. Used by permission of David DeVaus. Copyright # 1990. Surveys in Social Research. Allen and Unwin. Table 4.7. From Research Methods in Social Relations, sixth edition, by Charles M. Judd, Eliot R. Smith and Louise H. Kidder. Copyright # 1991, Holt, Rinehart and Winston. Reproduced by permission of the publishers. Every effort has been made to trace all copyright holders, but if any have been overlooked, or if any additional information can be given, the publishers will be pleased to make the necessary amendments at the first opportunity.

14. 1 Order at All Points Counting and accounting A man is driving through the bush one day and has to stop while a farmer takes his sheep across the road. There are quite a lot of sheep, so it takes a fair while. When they've all passed by, the man goes up to the farmer and asks, Ìf I can tell you how many sheep you have, to within one either way, can I have one of them?' The farmer replies, `Course you can. You'll never get it right.' The man says, `You have six thousand four hundred and twenty two.' `Well blow me down,' replies the farmer ± or words to that effect. Ìn fact I have six thousand four hundred and twenty one. I counted them this morning.' So the man walks back to the car with his prize. `Wait on,' cries the farmer. Ìf I can tell you what your job is, can I have her back?' `Sure,' says the man, `You'll never guess.' `Well,' says the farmer, Ì figure you'd be a statistician with the Australian Bureau of Statistics.' `Well I'll be . . . !' the man replies, Èxactly right. How on earth did you know that?' The farmer comes back: `Put me dog down and I'll tell you.' Traditional Australian Bush Yarn THE Ql-Qt CONTINUUM1 Like many in the humanities and social sciences, I was trained to be (at the least) sceptical about statistical methods and (at most) downright hostile towards them. In sceptical mode, I was exhorted to use statistics not in the way a drunk uses a lamppost: for support rather than illumination. In hostile mode, the word was that statistics was for `positivists' (a very unfair characterization, as it turns out, of positivism). What all of this well-mean- ing and humanistic advice ignored was the sheer fact that our social and cultural worlds, today, are massively subject to statistical accounts (see Hacking, 1982). Whenever we turn on the TV news or open a newspaper, the world is now routinely accounted for in terms of the numbers it gen- erates: from world population statistics right down to chewing gum mar- kets. In this respect, it's not quite as if numbers were on one side of the coin and `lived cultures' on the other. Rather, the technologies of numbering have become just one (though, in some disciplines, a dominant one) of the many practices that make up the cultures of modernity. In this brief introduction, then, I want to think through the supposed distinction (binary, even) between the quantitative (Qt) and the qualitative (Ql) and to show that the seal between the two is by no means as watertight as it is often assumed to be.

15. My first realization of an elision between Qt and Ql came to me when, out of sheer impecuniousness, I went to work for the Survey Research Centre at the Australian National University (ANU) in the mid-1970s. Prior to this way of supplementing my meagre PhD scholarship, my only encounter with statistics had been the compulsory undergraduate methods course in sociology, taught, as it happened, by a died-in-the-wool symbolic inter- actionist, a Ql-man if ever there was one! Said lecturer was, then, very happy for me to complete my statistics assignments by having a friend who was a physics student crunch the exercises on the university's one mainframe computer by submitting bundles of punchcards. Not, then, exactly the best of trainings or qualifications, I admit. But working late at ANU, designing and administering the Australian Capital Territory (ACT) population surveys, I came to see what a symbolic and interactional process Qt work could be in practice. One of our clients at the time was the local Family Planning organization. It wanted to know which forms of contra- ception were most in use in the Capital Territory. The only problem with this was that the official sampling procedures required interviewers to jointly interview two members of each household selected (using lot numbers) on a rotational basis: oldest and third oldest in odd-numbered lots, and second oldest and fourth oldest in even-numbered lots. This meant, in effect, that a fair proportion of interviews involved parents and their older children ± not exactly the best interactional setting to ask people about their contraceptive practices. The problem was both, and equally, statistical and `cultural.' Qt and Ql could not be a simple binary. And, oh yes, the wonder- ful `solution' we developed was to draw up a card with each kind of contra- ception numbered. Respondents were then shown the card and would say such things as `Well, I tried the number seven but it didn't work for me, so now I prefer the twenty six.' The same realization came back to me during a more recent research project (Mickler and McHoul, 1998). In this project, we collected over 600 newspaper articles on Aborigines, youth and crime over a 12-month period in the early 1990s in order to see whether there had been, as some suspected at the time, a media-generated `crime wave.' We had a neutral reader/ research-assistant type the articles into a relational database program (QSR NUDEIST) and, at the same time, code the articles for such things as `source' (the origin of the reported events), `participants' (the categories of persons reported on in each article) and how the reader thought the article was treating such `participants' (in positive, negative or neutral moral terms). What we hoped to get out of this was a strongly Ql argument based on a discursive analysis of the news articles and their `readings.' However, before long, we found that working with over 600 texts would not allow us to do this. The data in question were simply too numerous. And anyway, NUDEIST was starting to generate matrices of such things as `Date of publication' `Article source' and `Newspaper' `Participants.' Each cell of the matrix listed the relevant articles by their unique NUDEIST document number. There was no way we could work with this B A L N AV E S A N D C A P U T I 2

16. kind of output in a purely Ql mode. We had to show our findings graphically. This meant a lot of arduous work, transferring the totals of each cell into an Excel spreadsheet and then having Excel generate graphs and bar charts. (This was in NUDEIST version 3: the day we finished our gruelling transcription work, NUDEIST 4 was released, including new software that automatically converts such findings into popular spreadsheet and stats programs.) At that point we were able to generate `findings' such as those in Chart 1.1. This was very useful to us because it showed us a picture of the year's news in terms of just when the WA press was reporting youth crimes, and the months during which young Aboriginal people were the `participants' in those reports. However, while this could show us the media angle, it could not tell us whether or not the `peaks' of youth crime reportage (and we were particularly interested in the September peak because it followed the infamous Rally for Justice outside Parliament House in the August) corresponded to `actual' crime rates. If we were going to find a `media wave,' we would have to overlay crime stats on to our bar chart. We even- tually tracked these down at the University of Western Australia Crime Research Centre whose staff kindly gave us their raw figures for the 12-month period in question. Accordingly we could then generate more conclusive findings about the supposed `wave' ± see Chart 1.2. O R D E R AT A L L P O I N T S 0 5 10 15 20 25 30 35 FEB '91 MAR APR MAY JUNE JULY AUG SEP OCT NOV DEC JAN '92 Date in Months No. of News Reports Reports on Aboriginal Youth All Youth-Crime Reports CHART 1.1 Reports on Aboriginal youth and all youth-crime reports, Feb.1991to Jan.1992 3

17. So we now had something like a `media wave.' That is, as the reportage of youth crime peaked in September there was a corresponding decline in `actual' crimes as measured by apprehensions and reports to the police. Again, we could only see this graphically ± though later we resorted to a more Ql discussion of how, in terms of routine media practices, the `wave' could have been generated. That is, we wanted to show, from a close reading of the reports (in line with well-known facts about the news profession (Ericson et al., 1987, 1989, 1991)) that the statistically visible `wave' was not so much a conscious conspiracy as a routine phenomenon in the field of `bad news as usual.' As so frequently, the raw stats were open to more than one interpretation and only one of these was that the media were `deliber- ately' causing the appearance of a crime wave with possibly suspect and populist motives. Working on this project, then, brought me to the realization that there is in fact a continuum between hypothetically `pure' Qt and Ql analysis. The two, that is, never actually exist, empirically, in their pure states. Qt work is part of the `culture' at large and Ql work is never strictly free of numbers. So why then is there an effective stand-off between professionals in the two methodological fields? Why do some Qt workers insist on accepted sampling procedures and statistical reliability while some Ql workers B A L N AV E S A N D C A P U T I 0 10 20 30 40 50 60 70 FEB '91 MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC JAN '92 Date in Months No. of News Reports (10=10), No. of Apprehension (10=100) and No. of Offences Reported to Police (10=1000) Aboriginal Youth-Crime Reports All Youth-Crime Reports All Juvenile Apprehensions Aboriginal Juvenile (Male) Apprehensions Stolen Motor Vehicles, Break and Enter and Burglaries Reported to Police (Crime Statistics Sources: Broadhurst and Ferrante (1995: 87); Broadhurst and Loh (1995: 72)) CHART 1.2 News reports on Aboriginal youth-crime and all youth-crime with actual crime data, Feb.1991to Jan.1992 4

18. argue that very small fragments of cultural `data' (a poem, a few turns at talk, an advertisement, and so on) can generate just as fruitful findings? In order to begin to answer this question, we can turn to a rather extreme case of an anti-Qt argument. The one I have in mind was mounted some years ago by another major influence on my own methodological preferences: Harvey Sacks. CULTURE AND THE RELIABILITY OF `DATA'2 Just as in physics, natural phenomena must be massively sampled before an argument about their properties can be mounted, so, in some branches of the social sciences, cases must be sufficiently large in number before gen- eralizations can be acceptable to the professional community. The parallel is both noble and mistaken, according to several social theorists of (roughly) a phenomenological persuasion. Alfred Schu Ètz, many years ago, and in works that are almost unknown to Qt types today, put this argument very simply (Schu Ètz, 1962: 34-47). He noticed that natural scientists deal with objects of the first order of interpretation: that is, the objects that the natural scientist has before her have not been interpreted until she herself shows up on the scene. Atoms, planets and cyclones do not interpret themselves. However, as soon as we come to deal with human individuals or populations, we're in a different situation entirely. These òbjects of knowledge' have interpreted themselves before the social scientist comes on the scene. They are therefore òbjects' of the second order of interpretation. They come pre-interpreted ± and so the job of anyone working in the human sciences is to interpret the interpretations that people have already made of their lives (individually) and their cultures (collectively). Where human beings are concerned (as opposed to rats or atoms), the investigator is always investigating `second order' matters. It is a fundamental category mistake, then, to treat humanly produced data as `natural' data ± though, as we will see shortly, this logical error need not preclude Qt analysis altogether: I am simply taking this argument as one of the most cogent to date on the underlying `philosophy' of much Qt thinking. Sacks's important intervention into this debate ± possibly via his association with Garfinkel (1967) who was, in turn, majorly influenced by Schu Ètz ± is to argue for a unique position for cultural `data.' Not only, he ventures, is culture distinct from nature, it also has unique orderly properties. That is: in coming to analyse human cultures, one must take into account that they display òrder at all points.' The argument is complex and students would do best to consult the original (1995: 484). However, Schegloff (who introduces both volumes) puts the matter succinctly. Accordingly, I quote at length: Considerations of enculturation and `language acquisition' provide an especially provocative focus for a matter which Sacks raises . . . as a methodological point. O R D E R AT A L L P O I N T S 5

19. Taking up the relevance of sampling, Sacks points out that it depends on the sort of order one takes it that the social world exhibits. An alternative to the possibility that order manifests itself at an aggregate level and is statistical in character is what he terms the òrder at all points' view. . . . This view, rather like the `holographic' model of information distribution, understands order not to be present only at aggregate levels and therefore subject to an overall differential distribution, but to be present in detail on a case by case, environment by environment basis. A culture is not then to be found only by aggregating all of its venues; it is substantially present in each of its venues. (Schegloff in Sacks 1995: xlvi) As with holographs, a culture must display its inherent order, no matter how much or little of it we happen to examine or `sample.' (And, as we will see shortly, this importantly applies as much to the `much' as it does to the `little.') Even if 99 per cent of a holographic image is destroyed, the remain- ing 1 per cent contains all of the information contained in the original whole (Roe, 1998). In this respect, when we are dealing with social and cultural phenomena (such as `mental illness' or `suicide,' for example), the question of whether we have sufficient cases for the `data' to be `reliable' may not arise. The order we are seeking, as social scientists or cultural analysts, according to Sacks, is not available as a form of statistical aggregation; rather it must be inherently present in whatever materials happen to be at hand ± for example just a few turns in a conversation. And, in fact, Sacks directly uses his òrder at all points' (or ìn all venues') model to criticize Qt research. His argument here is particularly directed at sociology and survey research, but it applies equally well to the other social sciences: . . . you might find that the fact of òrder at all points' could be used to explain fairly strange facts; things like the following. For one, certain kinds of researches ± for example, let's say, taking very conventional sociology, survey research ± would use the fact that it gets orderly results to indicate that it must be doing something decent. Now everybody knows ± that is, everybody who has ever done such work, or studied it ± that it's almost universally extremely bad; that almost never can whatever the constraints are that ought to be used under their own formulation to decide that they have something, be used. There are, for example, all kinds of constraints on proper statistical procedures which are never satisfied, and none the less they get order. (1995: 484, my emphasis) The argument here is important: the reason why Qt procedures arrive at orderly outcomes may not be because of the `reliability' of either the sample or the statistical techniques deployed on that sample. Rather, the order that arises would be, for Sacks, an effect of the fact that, wherever a culture is tapped into, and however frequently or infrequently it is sampled, it will display its inherent order. If this is the case ± and I have yet to see a social theory that directly defends an aggregationist model against the idea of òrder at all points' ± then statistically reliable results may well be nothing more than a case of something intrinsically non-statistical showing through (and despite) the sampling techniques employed in any given case. The B A L N AV E S A N D C A P U T I 6

20. consequences of such a `paradigm shift' are potentially massive for social scientific thinking ± not least because they might change our thinking away from what Wittgenstein (1958b: 18) called our `craving for generality' and/ or `the contemptuous attitude towards the particular case' (cf. Turner, 1974: 213). However, none of this means that Qt work is without value.3 On the contrary, as we have seen in our first section of this introduction, statistical representations are sometimes inevitable: and in the above case, the reason for this had to do with the necessarily longitudinal nature of the `media wave' hypothesis. If this is so ± and because Ql and Qt exist on a continuum and not as strict alternatives ± then even the most Ql-oriented researchers will occasionally find themselves needing to use statistical models and their ensuing graphic representations. If the `continuum' argument holds up (and I think it does, at least from experience), then how can this live in the same world as the òrder at all points' argument? Sacks himself gives us the answer when he writes that, even when we do proceed statistically, and even in its crudest form (the survey), we will find the same cultural order as when we do not! `None the less they get order' is precisely what he tells us. That is, if cultural `data' is intrinsically data of the second order, any statistical analysis of it cannot help but preserve that aspect of it. The `philosophy' of treating social and cultural phenomena as if they were natural phenomena may well be a philosophical error ± associated (probably wrongly) with `positivism.' But this does not mean that the methods that flow from that error can somehow delete the orderly cultural properties of the former kind of phenomenon. In fact, Sacks himself tells us that they could not even if they wanted to. This then sets the tone ± or more precisely, the ànalytic mentality' (Schenkein, 1978) ± for the following chapters of this book.4 They introduce some central social and cultural uses of Qt methods and, at the same time, always remember the importance of the òrder at all points' that such uses must bring out. Or to put this another way: there is no logical reason why Qt methods can't be just as informed by a Sacksian view of cultural order as they can by starting `positivistically' with physics envy. The result is, I think, the first textbook in the field to bring cultural responsiveness to statistical methods and, equally, statistical responsibility to cultural analysis. Alec McHoul Murdoch University 6th January 2000 NOTES 1. This sub-title with apologies to Star Trek TNG. In many way, Star Trek's Q is the ultimate detective. His case is nothing more or less than the human condition. O R D E R AT A L L P O I N T S 7

21. 2. This section is based on parts of an article on Sacks, conversation analysis and psychology by Alec McHoul and Mark Rapley (2000). For a good general introduction to Sacks, see Silverman (1998). 3. For example, the foundational text of ethnomethodology, Garfinkel's Studies in Ethnomethodology (1967), extensively uses statistical and quasi-statistical representations to show èxperimental' results. 4. I write `more precisely' with half an eye on the problem of what `precision' could mean on the Ql±Qt continuum. To explain: the term ànalytic mentality' is from Jim Schenkein (1978). He writes: Ì am never entirely sure of what I mean by `ànalytic mentality'', but then goes on to say: Ì have come to think of [our] shared practices [in conversation analysis] as ingredients in an `ànalytic mentality'' since other more familiar terms fail to capture the spectrum of ingredients sketched here. Our particular grouping of shared dispositions is not exactly a ``theory'' or ``method'', nor is it merely a ``point of view'' or some kind of ``philosophy''. But whatever it is, the result is a research environment . . . and `ànalytic mentality'' is fuzzy enough to embrace whatever comes up when a closer look at the environment is taken' (1978: 6). So this fuzzy formulation seems to be, paradoxically perhaps, a more precise description than terms like `theory' and `method'. This would be exactly my characterization of work that explicitly acknowledges and works along the Ql±Qt continuum. As Wittgenstein (1958a: #71) puts it: Ìs it . . . always an advantage to replace an indistinct picture by a sharp one? Isn't the indistinct one often exactly what we need?' REFERENCES Broadhurst, Roderic and Ferrante, Anna (1995) `Trends in juvenile crime and justice, 1990±1992', in R. Harding (ed.) Repeat Juvenile Offenders: The Failure of Selective Incapacitation in Western Australia. Nedlands, WA: Crime Research Centre, University of Western Australia. pp. 79±95. Broadhurst, Roderic and Loh, Nini Sui Nie (1995) `Selective incapacitation and the phantom of deterrence', in R. Harding (ed.) Repeat Juvenile Offenders: The Failure of Selective Incapacitation in Western Australia. Nedlands, WA: Crime Research Centre, University of Western Australia. pp. 55±78. Ericson, Richard, Baranek, Patricia and Chan, Janet (1987) Visualizing Deviance: A Study of News Organization. Toronto: University of Toronto Press. Ericson, Richard, Baranek, Patricia and Chan, Janet (1989) Negotiating Control: A Study of News Sources. Stony Stratford: Open University Press. Ericson, Richard, Baranek, Patricia and Chan, Janet (1991) Representing Order: Crime, Law, and Justice in the News Media. Buckingham: Open University Press. Garfinkel, Harold (1967) Studies in Ethnomethodology. Englewood Cliffs: Prentice- Hall. Hacking, Ian (1982) `Biopower and the avalanche of printed numbers', Humanities in Society, 5 (3/4): pp. 279±95. McHoul, Alec and Rapley, Mark (2000) `Sacks and clinical psychology', Clinical Psychology Forum, 142 (August): pp. 5±12. Mickler, Steve and McHoul, Alec (1998) `Sourcing the wave: Crime reporting, Aboriginal youth and the WA Press, Feb 1991±Jan 1992', Media International Australia incorporating Culture and Policy, 86 (February): pp. 122±52. B A L N AV E S A N D C A P U T I 8

22. Roe, Phil (1998) Of Hologrammatology: The Politics of Virtual Writing. Unpublished PhD thesis, Murdoch University. Sacks, Harvey (1995) Lectures on Conversation, Vol. 1. Ed. G. Jefferson, intro. E.A. Schegloff. Oxford: Blackwell. Schenkein, Jim (1978) `Sketch of an analytic mentality for the study of conversational interaction', In J. Schenkein (ed.) Studies in the Organization of Conversational Interaction. New York: Academic Press. pp. 1±6. Schu Ètz, Alfred (1962) Collected Papers Vol. 1: The Problem of Social Reality. Ed. and intro. M. Natanson. The Hague: Martinus Nijhoff. Silverman, David (1998) Harvey Sacks: Social Science and Conversation Analysis. Cambridge: Polity Press. Turner, Roy (1974) `Words, utterances and activities', in R. Turner (ed.) Ethnomethodology: Selected Readings. London: Penguin. pp. 197±215. Wittgenstein, Ludwig (1958a) Philosophical Investigations. 2nd edn. Trans. G.E.M. Anscombe. Oxford: Blackwell. Wittgenstein, Ludwig (1958b) The Blue and Brown Books: Preliminary Studies for the `Philosophical Investigations'. Oxford: Blackwell. O R D E R AT A L L P O I N T S 9

23. 2 Starting the Inquiry `But what happened then?' `I'll tell you what happened then,' said Fiennes with a grim emphasis. `When we got back into that garden the first thing we saw was Traill, the lawyer; I can see him now with his black hat and black whiskers relieved against the perspective of the blue flowers stretching down to the summer-house, with the sunset and the strange outline of the Rock of Fortune in the distance. His face and figure were in shadow against the sunset; but I swear the white teeth were showing in his head and he was smiling. `The moment Nox saw that man the dog dashed forward and stood in the middle of the path barking at him madly, murderously, volleying out curses that were almost verbal in their dreadful distinctness of hatred. And the man doubled up and fled along the path between the flowers.' The Oracle of the Dog Fiennes was a friend of Father Brown, G.K. Chesterton's famous priest- detective. Fiennes was recounting a story to Father Brown about the reaction of Nox, the dog, to a lawyer who represented Colonel Druce, a wealthy man who had only recently been murdered. The newspapers had reported that Druce had been murdered with a sharp knife in a garden with no outside entry. The witnesses saw no murderer and there was no knife. Fiennes thought that the dog's reaction to Traill the lawyer was a clear sign that the lawyer was the murderer. Father Brown, however, interrogates his friend further and discovers that Colonel Druce changed his will and left his money to his daughter, rather than to his son Harry Druce, who owed money from gambling. Harry Druce committed suicide after his father's death. Father Brown also found out that the dog had tried to retrieve something from the ocean on a walk with Fiennes at the time of the murder and came back empty handed. Father Brown comes to the conclusion that Colonel Druce, who wore a white coat, was visible through a hedge that surrounded the garden. A man with a walking-stick-knife had seen Druce through the hedge and stabbed him. The man had then thrown the walking-stick-knife into the ocean. The dog tried to retrieve the walking stick, but to no avail because it sank. Harry Druce was the man who killed Colonel Druce. Harry killed himself when he realized that his murder was in vain and that his severe debts would not be resolved. Fiennes is amazed that Father Brown can work out the crime from a distance. Father Brown replies, `You asked how I could guess things a hundred miles away but honestly it's mostly to your credit, for you

24. described people so well that I know the types.' Father Brown rejected Fiennes's interpretation of the dog's reaction to Traill and instead broad- ened the inquiry. Ì had a sort of guess,' said Father Brown, `right at the beginning when you said that Druce wore a white coat.' Father Brown was not a Columbo or a Hercule Poirot, who often as not started their investigations knowing who the crook was. Father Brown started his inquiries into mysterious unsolved crimes by collecting a wide range of accounts and observations, no matter how zany or far fetched those accounts and observations might superficially look. Father Brown did not assume that the context of a crime was obvious and that knowledge of that context came from looking at only a few clues. Father Brown carefully weighs up what he has found (what he thinks is a clue) and the logical consequences of what he has found (what he thinks must rationally follow). In The Strange Crime of John Boulnois (Chesterton, 1987), Father Brown arrives at a scene where an American journalist has witnessed the death of Sir Claude Champion, dressed as Romeo and stabbed with a sword, saying with his last breath `Boulnois . . . with my own sword . . . he threw it . . .'. The journalist was on his way to interview John Boulnois, an Òxford man' who had recently published a review on Darwinian evolution. Father Brown interviews the wife of John Boulnois. `Father Brown' she said. `Mrs. Boulnois?' he replied gravely. Then he looked at her and immediately said: Ì see you know about Sir Claude.' `How do you know I know?' she asked steadily. [emphasis added] He did not answer the question, but asked another: `Have you seen your husband?' `My husband is at home,' she said. `He has nothing to do with this.' Again he did not answer; and the woman drew nearer to him, with a curiously intense expression on her face. `Shall I tell you something more?' she said, with a rather fearful smile. Ì don't think he did it, and you don't either.' Father Brown returned her gaze with a long, grave stare, and then nodded, yet more gravely. `Father Brown,' said the lady, Ì am going to tell you all I know, but I want you to do me a favour first. Will you tell me why you haven't jumped to the conclusion of poor John's guilt, as all the rest have done? Don't mind what you say: I±I know about the gossip and the appearances that are against him.' Father Brown looked honestly embarrassed and passed his hand across his fore- head. `Two very little things,' he said. Àt least one's very trivial and the other very vague. But such as they are, they don't fit in with Mr. Boulnois being the murderer.' He turned his blank, round face up to the stars and continued absent mindedly: `To take the vague idea first. I attach a good deal of importance to vague ideas. All those things that àren't evidence' are what convince me. [emphasis added] I think a moral impossibility the biggest of all impossibilities. I know your husband only slightly, but I think this crime of his, as generally conceived, something very like a moral impossibility. Please do not think I mean that Boulnois could not be so S TA R T I N G T H E I N Q U I R Y 1 1

25. wicked. Anybody can be wicked ± as wicked as he chooses. We can direct our moral wills; but we can't generally change our instinctive tastes and ways of doing things. Boulnois might commit a murder, but not this murder. He would not snatch Romeo's sword from its romantic scabbard; or slay his foe on the sundial as on a kind of altar; or leave his body among the roses; or fling the sword away among the pines. If Boulnois killed anyone he'd do it quietly and heavily, as he'd do any other doubtful thing ± take a tenth glass of port, or read a loose Greek poet. No, the romantic setting is not like Boulnois. It's more like Champion.' Àh!' she said, and looked at him with eyes like diamonds. Ànd the trivial thing was this,' said Brown. `There were fingerprints on that sword; fingerprints can be detected quite a time after they are made if they're on some polished surface like glass or steel. These were on a polished surface. They were half-way down the blade of the sword. Whose prints they were I have no earthly clue; but why should anybody hold a sword half-way down? It was a long sword, but length is an advantage in lunging at an enemy. At least, at most enemies. At all enemies except one.' Èxcept one!' she repeated. `There is only one enemy,' said Father Brown, `whom it is easier to kill with a dagger than a sword.' Ì know,' said the woman. Òneself.' There was a long silence, and then the priest said quietly but abruptly: Àm I right, then? Did Sir Claude kill himself?' `Yes,' she said, with a face like marble. Ì saw him do it.' `He died,' said Father Brown, `for love of you?' An extraordinary expression flashed across her face, very different from pity, modesty, remorse, or anything her companion had expected: her voice became suddenly strong and full. Ì don't believe,' she said, `he ever cared about me a rap. He hated my husband.' `Why?' asked the other, and turned his round face from the sky to the lady. `He hated my husband because . . . it is so strange I hardly know how to say it . . . because . . .' `Yes?' said Brown patiently. `Because my husband wouldn't hate him.' Father Brown only nodded, and seemed still to be listening; he differed from most detectives in fact and fiction in a small point ± he never pretended not to under- stand when he understood perfectly well. Mrs. Boulnois drew near once more with the same contained glow of certainty. `My husband,' she said, ìs a great man. Sir Claude Champion was not a great man: he was a celebrated and successful man. My husband has never been celebrated or successful; and it is the solemn truth that he has never dreamed of being so. He no more expected to be famous for thinking than for smoking cigars. On all that side he has a sort of splendid stupidity. He has never grown up. He still liked Champion exactly as he liked him at school; he admired him as he would admire a conjuring trick done at the dinner-table. But he couldn't be got to conceive the notion of envying Champion. And Champion wanted to be envied. He went mad and killed himself for that.' (Used by permission) Father Brown's detection style is summarized in his enigmatic comment Àll those things that `àren't evidence'' are what convince me'. Father B A L N AV E S A N D C A P U T I 12

26. Brown tries to avoid preconceived hypotheses and preconceived ideas about the relevance of the things that he observes. Father Brown works from ìnside out', carefully analysing all the events and clues (including witness accounts) and then drawing conclusions based on the evidence. In many of the Father Brown stories, of course, there is something Father Brown knows that we do not know. We only find out at the end of the story how Father Brown reached his conclusions and why. The detective's knowledge and the reader's knowledge are not always the same during a story. For Father Brown and the reader, however, there are only a finite number of interpretations of accounts available to explain the crime mystery. Indeed, we would not want to read detective fiction if we thought that any story accounted for what had happened. If Father Brown works from the ìnside out' in his collection and analysis of evidence, then Sherlock Holmes is perhaps the perfect example of the detective who works from the òutside in'. Father Brown, like Holmes, always held that there is a rational explanation for all things. Father Brown, though, is seen as the inquisitive interviewer, collecting individual facts and then solving the crime. Sherlock Holmes, in contrast, argued that he could deduce the solution to the whole crime from only a few facts. Holmes always enjoyed demonstrating his `powers of deduction' with Dr Watson. Dr Watson recounts this scene from A Scandal in Bohemia. One night ± it was on the twentieth of March, 1888ÐI was returning from a journey to a patient (for I had now returned to civil practice), when my way led me through Baker Street. As I passed the well-remembered door, which must always be associated in my mind with my wooing, and with the dark incidents of The Study in Scarlet, I was seized with a keen desire to see Holmes again, and to know how he was employing his extraordinary powers. His rooms were bril- liantly lit, and, even as I looked up, I saw his tall, spare figure pass twice in a dark silhouette against the blind. He was pacing the room swiftly, eagerly, with his head sunk upon his chest and his hands clasped behind him. To me, who knew his every mood and habit, his attitude and manner told their own story. He was at work again. He had risen out of his drug-created dreams and was hot upon the scent of some new problem. I rang the bell and was shown up to the chamber which had formerly been in part my own. His manner was not effusive. It seldom was; but he was glad, I think, to see me. With hardly a word spoken, but with a kindly eye, he waved me to an armchair, threw across his case of cigars, and indicated a spirit case and a gasogene in the corner. Then he stood before the fire and looked me over in his singular intro- spective fashion. `Wedlock suits you,' he remarked. Ì think, Watson, that you have put on seven and a half pounds since I saw you.' `Seven!' I answered. Ìndeed, I should have thought a little more. Just a trifle more, I fancy, Watson. And in practice again, I observe. You did not tell me that you intended to go into harness.' `Then, how do you know?' S TA R T I N G T H E I N Q U I R Y 13

27. Ì see it, I deduce it. [emphasis added] How do I know that you have been getting yourself very wet lately, and that you have a most clumsy and careless servant girl?' `My dear Holmes,' said I, `this is too much. You would certainly have been burned, had you lived a few centuries ago. It is true that I had a country walk on Thursday and came home in a dreadful mess, but as I have changed my clothes I can't imagine how you deduce it. As to Mary Jane, she is incorrigible, and my wife has given her notice, but there, again, I fail to see how you work it out.' He chuckled to himself and rubbed his long, nervous hands together. Ìt is simplicity itself,' said he; `my eyes tell me that on the inside of your left shoe, just where the firelight strikes it, the leather is scored by six almost parallel cuts. Obviously they have been caused by someone who has very carelessly scraped round the edges of the sole in order to remove crusted mud from it. Hence, you see, my double deduction that you had been out in vile weather, and that you had a particularly malignant bootslitting specimen of the London slavey. As to your practice, if a gentleman walks into my rooms smelling of iodoform, with a black mark of nitrate of silver upon his right forefinger, and a bulge on the right side of his top-hat to show where he has secreted his stethoscope, I must be dull, indeed, if I do not pronounce him to be an active member of the medical profession.' I could not help laughing at the ease with which he explained his process of deduction. `When I hear you give your reasons,' I remarked, `the thing always appears to me to be so ridiculously simple that I could easily do it myself, though at each successive instance of your reasoning I am baffled until you explain your process. And yet I believe that my eyes are as good as yours.' [emphasis added] `Then how do you know?', Watson asks Holmes. Ì see it, I deduce it', says Holmes in reply. What made Holmes's explanations the superior ones, in Watson's eyes, was the fact that Holmes could, in an apparently scientific and law-like way, deduce so much about events from so few clues. Sherlock Holmes and Father Brown differ in their detection styles but both have to `give reasons' for their explanations. `Then how did you know?' is a common refrain in detective fiction. Readers of detective fiction would not be happy if their favourite detective gave conclusions without reasons ± without justifications for what constituted evidence and how evidence and conclusions fitted together. But are there other styles of detection ± Flashes of insight? Sudden guesses? Intuition? With no link to evidence at all? Dirk Gently from Douglas Adams's Dirk Gently's Holistic Detective Agency is a good example of a detective who is not interested in the accounts people give him or in logical causal explanations of events. Dirk Gently's primary concern is with the `fundamental interconnectedness of all things' ± thus the `holistic' in the title of Gently's detective agency: `We solve the whole crime. We find the whole person' (Adams, 1987: 111). Dirk Gently tries to solve crimes with no reference to the logical connection between events, as in this scene with his secretary Miss Pearce. He thrust a piece of paper across the desk. She picked it up and looked at it. Then she turned it round and looked at it again. B A L N AV E S A N D C A P U T I 14

28. She looked at the other side and then she put it down. `Well?' demanded Dirk. `What do you make of it? Tell me!' Miss Pearce sighed. Ìt's a lot of meaningless squiggles done in blue felt tip on a piece of typing paper,' she said. Ìt looks like you did them yourself.' `No!' barked Dirk, `Well, yes,' he admitted, `but only because I believe that is the answer to the problem!' `What problem?' `The problem,' insisted Dirk, slapping the table, òf the conjuring trick! I told you!' `Yes, Mr Gently, several times. I think it was just a conjuring trick. You see them on the telly.' `With this difference ± that this one was completely impossible!' `Couldn't have been impossible or he wouldn't have done it. Stands to reason.' Èxactly!' said Dirk excitedly. Èxactly! Miss Pearce, you are a lady of rare perception and insight.' `Thank you, sir, can I go now?' `Wait! I haven't finished yet! Not by a long way, not by a bucketful! You have demonstrated to me the depth of your perception and insight, allow me to demon- strate mine!' Miss Pearce slumped patiently in her seat. Ì think,' said Dirk, `you will be impressed. Consider this. An intractable problem. In trying to find the solution to it I was going round and round in little circles in my mind, over and over the same maddening things. Clearly I wasn't going to be able to think of anything else until I had the answer, but equally clearly I would have to think of something else if I was ever going to get the answer. How to break this circle? Ask me how.' `How?' said Miss Pearce obediently, but without enthusiasm. `By writing down what the answer is!' exclaimed Dirk. Ànd here it is!' he slapped the piece of paper triumphantly and sat back with a satisfied smile. Miss Pearce looked at it dumbly. `With the result,' continued Dirk, `that I am now able to turn my mind to fresh and intriguing problems, like, for instance . . .' He took the piece of paper, covered with its aimless squiggles and doodlings, and held it up to her. `What language,' he said in a low, dark voice, ìs this written in?' Miss Pearce continued to look at it dumbly. Dirk flung the piece of paper down, put his feet up on the table, and threw his head back with his hands behind it. `You see what I have done?' he asked the ceiling, which seemed to flinch slightly at being yanked so suddenly into the conversation. Ì have transformed the problem from an intractably difficult and possibly quite insoluble conundrum into a mere linguistic puzzle. Àlbeit', he muttered, after a long moment of silent pondering, àn intractably difficult and possibly insoluble one.' He swung back to gaze intently at Janice Pearce. `Go on,' he urged, `say that it's insane ± but it might just work!' Janice Pearce cleared her throat. Ìt's insane,' she said, `trust me.' (Used by permission) S TA R T I N G T H E I N Q U I R Y 15

29. Dirk Gently sees a mystery where others do not. He provides meaningless answers to that mystery. His secretary does not accept that there is a mystery or the need for a solution. Ìt's insane', says Miss Pearce. This is of course a part of the humour of this novel. Gently is an exaggerated example of the role that `gut feelings', `guessing' and `lateral thinking' play in detective fiction. Guessing plays an important part in social science research, as we will see later. Dirk Gently, like Sherlock Holmes, believes that it is possible to explain events by reference to `wholes' ± to interconnected laws that govern ± on the surface ± seemingly unrelated events. Gently, like Holmes, held that if you knew one link in the causal chain then you could find the `whole'. This model of detection is similar to the `nomothetic' model in social science research. A nomothetic model is a macro, probabilistic, approach to analysing what happens in society. The aim in the nomothetic model is to identify general classes of actions or events in society and not to show all the individual unique events that may lie behind them. The nomothetic model of research is often contrasted with idiographic models in social science, which analyse all the micro events behind a social phenomenon (Babbie, 1986: 53±55). Father Brown's inquiries are idiographic in nature. He looks at all the micro events associated with a murder mystery. Both models of inquiry ± nomothetic and idiographic ± seek to explain what is happening. Both models must be grounded in èveryday life'. The art of the detective and the social scientist, of course, is in the ability to identify important events and to explain the relationships between them. Father Brown's question `But what happened then?' is the sign that an inquiry ± an investigation ± has started. A good detective and a good social scientist needs to know what, who, how and when to investigate. A good social scientist needs a research design. KNOWING WHAT TO RESEARCH Exploration, Description and Explanation In the murder mystery Cause of Death Dr Kay Scarpetta, a medical examiner `detective', is called out to a strange drowning in a naval shipyard for decommissioned naval ships and submarines. The first reaction of those investigating the incident was that it was simply a drowning: `He probably just drowned,' Green was saying. Àlmost every diving death I've seen was a drowning. You die in water as shallow as this, that's what it's going to be' (Cornwell, 1997: 14). Scarpetta, however, avoids `closure' ± premature finishing of the case. She takes the unusual step of diving to the body to investigate what has happened ± to satisfy her own curiosity about what happened. She finds the air hose tangled on the side of an old ship but decides that the diver could have rectified this problem relatively easily. Scarpetta, after conducting the B A L N AV E S A N D C A P U T I 16

30. postmortem, finds that there is also cyanide in the body. The murder mystery gets more and more complex as the novel progresses, leading ulti- mately to a major conspiracy to take over a nuclear power plant. Scarpetta explores, describes and then explains what is happening. Social science research sometimes involves all three kinds of research and sometimes only one. Exploratory research in social science is valuable when a researcher wants to study a new area and/or to test methods, such as surveys and survey questions, for investigating that area. Descriptive research is one of the most common forms of research in social science research. The census and surveys of public opinion are examples of descriptive research. The census, for example, provides an overview of demographics (e.g. gender, income) of a whole population. Public opinion polls show what people's voting intentions might be. A descriptive study may raise issues that need explaining. Explanatory research, however, reports not just `what is happening' but `why'. Let's look briefly at some descriptive and explanatory studies on the sociology of news and the journalism profession. In January 1993 John Henningham published preliminary findings from a national survey of 1,068 Australian journalists, The Hack's Progress (1993: 45). The composite picture of a journalist, he said, is very different from the stereotype of middle-aged dissipation. The closer image is a 1990s yuppie. The survey shows that the typical journalist is male, young, and ambitious, with a middle-class, Anglo-Saxon background. There is an increasing chance that the journalist had a tertiary education and is likely to be well paid, content, committed to the career and optimistic about the future of the media. . . . About 37 per cent say they are more likely to vote Labor, with 29 per cent Liberal and 2 per cent National. The lean to the left is not a matter of background: most journalists come from middle-class families, with fewer than three out of 10 from blue-collar homes. Only one in 10 had parents in primary industry ± cause for rural people to continue their complaints of limited and supposedly unsym- pathetic coverage of their problems. The churches might also argue under repre- sentation, with 74 per cent of journalists uninvolved in religion. Only 19 per cent of journalists were born overseas and almost all of them came from an Anglo- Celtic background. Fewer than 3 per cent of journalists are non-Caucasian. Women hold one in three jobs, but this is an advance on the early 1970s, when the ratio was one-in-10. Predictably, while 72 per cent of women say it is harder for females to advance their careers, only 39 per cent of male journalists agree. Women have an average of 27, against 37 for men. The combined median is 32, young enough to explain some of the optimism found in the survey. The most extensive national survey of journalists ever undertaken is that done by Johnstone, Slawski and Bowman in the United States in 1971 and published in 1976. They surveyed 1,313 journalists by telephone from a national sample of 1,550. The authors estimate that fewer than half (45.7 per cent) of the 153,000 persons who reported themselves as `editors or S TA R T I N G T H E I N Q U I R Y 17

31. reporters' in the 1970 census were employed full-time in the American news media. They estimated the total full-time editorial manpower in English- language news media in the United States at 69,500 and said that three- quarters of them were employed in print media, about 20 per cent in broad- cast media and about 5 per cent in wire services at the time of their survey. More than half were employed by daily newspapers. Although they estimated that freelancers might contribute up to 100,000 stories a week to the news media, they did not include freelancers. Johnstone, Slawski and Bowman found their 70,000 US journalists to be overwhelmingly young, urban, mobile and male ± just as in Australia. They also found that their journalists did not conform to the stereotype of hard- drinking, callous and isolated creeps. In fact, they came from the same social strata as those in charge of the economic and political systems. There were few recruits from the working class. Journalists did not have to have a degree to be a journalist but in fact 86 per cent of US journalists had attended college for one or more years, about 60 per cent were graduates and more than 18 per cent had done postgrad- uate work. Although most studied journalism more than any other specific subject, journalism majors were outnumbered by about 2 to 1 by persons with other kinds of college training. The other subjects most in demand were political science and government. Johnstone, Slawski and Bowman's national survey of 1971 was replicated in 1982±3 by David Weaver and Cleveland Wilhoit (1986). In the years between 1971 and 1983 the USA had been involved in the Vietnam War, All The President's Men had exposed the criminal behaviour of Nixon and his aides in the White House, films such as Absence of Malice had questioned the ethics of journalists, the American media had been faced with a series of astronomical libel judgements, Janet Cooke had been awarded a Pulitzer Prize for a Washington Post story which she had faked and public confidence in the press had been eroded. Nevertheless, there was a 125 per cent increase in journalism and mass communication enrolments between 1971 and 1985. By 1983 there were an estimated 112,072 journalists in the United States. Weaver and Wilhoit surveyed 1,001 of them. They found the typical journalist to be a politically moderate, 32-year-old, college-educated, white Protestant who earned $19,000 a year. The proportion of women had increased from 20 per cent in 1971 to 34 per cent in 1983. Two out of three journalists saw themselves as middle-of-the-road politically, 22 per cent saw themselves as left of centre and 18 per cent to the right. This was interesting because shortly before the Weaver and Wilhoit study was published, Lichter, Rothman and Lichter published their survey of `elite' journalists ± those working on 10 major news outlets ± and said one out of two were left-of-centre, if the centre is defined as American businessmen. Their journalists were more likely than the national sample to see their role as challenging government. The Weaver and Wilhoit national sample were B A L N AV E S A N D C A P U T I 18

32. more moderate: most saw their role as being `interpretive'. The next biggest proportion saw their role as disseminating information. Only a small proportion espoused the adversary role. In 1971, 58 per cent were college graduates whereas in 1983, 70 per cent were and more than half of those majored in journalism or communications. Generally, journalists earned less in real dollars than they did in 1971. Most said they heard more about their work from readers and viewers than from colleagues. When they studied another sample of 1,400 journalists in the 1990s they found that journalists had less autonomy than reported in the earlier studies and less job satisfac- tion. The proportion of journalists planning to leave the profession was double that of the 1983 studies. But the typical American journalist remained a married white male in his thirties, with a bachelor's degree (Weaver and Wilhoit, 1996). Similar findings emerge from the British studies. In his Journalism Recruitment and Training: Problems in Professionalization Oliver Boyd-Barrett reports a 1969 survey of 99 trainee journalists in the two largest training centres (1970: 181±201). More than half of them came from lower-middle and upper-middle white-collar backgrounds and nine came from senior executive±managerial±professional backgrounds. About half could qualify for entrance to a university and he cites estimates that graduates comprise from 6 to 15 per cent of the total intake of recruits to journalism each year. Most (62 per cent) of those in his sample who were employed by newspapers obtained their first job by writing letters on their own initiative to one or several newspapers and another 25 per cent answered press adver- tisements. Boyd-Barrett's work also had an explanatory element. He investigated the reasons for choosing journalism as a career: it was seen as a non-routine, non-conventional, sociable occupation by 35 per cent of respondents; it was the most desirable occupation available to 29 per cent; it was seen as creative by 16 per cent and more than 75 per cent wanted to write a book (a novel in most cases); it was seen as self-educational by 5 per cent; as a `bridging occupation' to a better job by 3 per cent; and as a public service occupation by 1 per cent. In spite of the aspirations for the non-routine, non- conventional and the creative, in practice the young journalists spent a quarter or more of their work time at desk work or office work and most of their outside work on covering routine, predictable events. Feature writing accounted for only 10 per cent of total time. Nevertheless, they felt little dissatisfaction with the organization of their work or its opportunities. Most of them wanted to go into feature writing, general reporting or special writing which they saw as offering more opportunities for self-expression and initiative. Boyd-Barrett (1970: 60±4) says that most British recruits to journalism begin on weekly provincial newspapers and that juniors made up about two-fifths of the journalists on these papers in each of the years 1964±7. Almost all start as general reporters and their first chance to specialize is usually in sport. Although more than 60 per cent of national newspaper S TA R T I N G T H E I N Q U I R Y 19

33. journalists in Britain were over 35, the belief was widely held that `on your 35th birthday they push you down the rubbish chute'. It was also widely asserted that many journalists were sacked but Tunstall (1970) could find little evidence of this for the period 1965±8. An understanding of the demographics of the journalism profession as well as journalists' motivations for joining the profession combines descriptive and explanatory aspects of social science research. The `what' of the research and the `why' of the research overlap. In the surveys cited above it is possible to gain an overview of who journalists are and to discover trends in the journalism profession, including journalists' attitudes towards journalism. Topics of Research Knowing what to research, knowing the purpose of the research, is key to the first steps in a research design. Knowing whether the research is exploratory, descriptive or explanatory is a part of knowing what to research. The topics of social science research range from description ± describing who the people or social groups are ± to explanation ± explaining why those people or social groups think or behave the way they do. Analysing how people or social groups think and behave includes analysis of attitudes, beliefs, personality traits and other factors associated with motivation. Analysing how people or social groups behave includes analysis of the actual actions. Table 2.1, for example, tells us what is happening with Western Samoan television behaviour, but it does not tell us why the Samoans changed their behaviour. Why did they stop weaving mats? Was television the cause? What people do and what they say they do can also differ to a great degree. In 1934 Lapiere travelled with a Chinese couple to 66 hotels, auto- B A L N AV E S A N D C A P U T I TABLE 2.1 Impact of television onWestern Samoan evening activities, 4^10 pm % of time spent Without TV With TV Prayers 6 4 Eating 17 10 Talking/Discussion 30 9 Story telling/Singing 15 3 Visiting friends/ Community activities 22 4 Playing cards/ Weaving mats 10 0 WatchingTV 0 70 Source: Martin (1987: 3^21) 20

34. mobile camps, tourist homes, and 184 restaurants and cafes in the United States. The Chinese couple were accepted with open arms. Lapiere then sent questionnaires to all the places he visited, asking them if they would accept Chinese as guests. Ninety per cent replied `No' (Deutscher, 1973). Detectives like the medical examiner Kay Scarpetta go through all the steps of modern social science research. There is an exploratory element ± seeing whether there is a suspicious death; there is a descriptive element ± describing what actually happened to the dead person; and there is an explanatory element ± the reasons for the person's murder by cyanide. Scarpetta is interested in what people tell her but she is also interested in the actual behaviour of the people she talks to. Knowing what to research involves decisions on exploration, description and explanation as well as decisions about the topic of the research itself. These decisions will form part of a research design. Knowing whom to research is also an important part of research design. Knowing whom is not simply a matter of which people might be relevant to a study, but which units of analysis. KNOWING WHOM TO RESEARCH Units of Analysis Surveys on journalism as a profession involve individual journalists as the unit of analysis ± each journalist is asked a question in a survey. The individual journalist is also a member of a larger grouping ± the profession itself. If we make the explanatory statement `educated men are more likely than educated women to be journalists', then we are working at the level of the individual as the unit of analysis. If we make the descriptive statement that `the typical American journalist is a married white male in his thirties, with a bachelor's degree', then we are also dealing with the individual as the unit of analysis. However, if we said that journalists are more likely than accountants to be tertiary educated, then we would be using the group as the unit of analysis. The social science researcher also has to decide who counts as a `journalist'. This is especially important if we are drawing a `sample' of journalists for investigation. We will deal with samples and populations later, but there are important issues associated with the definition of a `journalist'. Is a journalist anyone who belongs to the journalists' association or is it anyone who writes a story for a publication (a local newsletter?). Is anyone who runs their own website a journalist? The `who' of research involves identifying the individuals who might be relevant to a study and the social groups to which they belong. Sherlock Holmes did this in his own way when he told Watson how he identified him. `As to your practice, if a gentleman walks into my rooms smelling of iodoform, with a black mark of nitrate of silver upon his right forefinger, S TA R T I N G T H E I N Q U I R Y 21

35. and a bulge on the right side of his top-hat to show where he has secreted his stethoscope, I must be dull, indeed, if I do not pronounce him to be an active member of the medical profession.' In social science research it is essential that you identify what your unit of analysis is. Is it criminals or crimes? Is it journalists or journalism? Understanding your units of analysis forms part of your decisions on research design. Poor definition of the unit of analysis and poor selection of those units raise questions about the external validity of a study ± can the results be generalized from a sample to all the units of analysis you are studying? Your literature review will also help you to clarify the what and the who of your study. KNOWING HOW TO RESEARCH Literature Review Coming up with good topics in social science research is not easy. Narrowing the topic ± giving it a decent focus and decent definition ± is even more difficult. One of the most complex research topics this century has been `Does media violence cause violent behaviour?' Popular concern about the relationship between media and behaviour has been ongoing. In 1993, international news agencies reported the murder of a four±year- old English boy, James Bulger, by two 10-year-olds who kidnapped him from a shopping centre and took him to a railway line. The four- year-old suffered 30 blows inflicting 22 injuries to his head, another 20 to his body. Bags containing 27 bloodstained bricks were also shown to the jury together with a 22 kg railway bolt used in the attack. The 10- year-old boys had watched over 300 adult violent videos in the previous six months. News media and the popular imagination often put media violence and violent behaviour together in a causal relationship. But a causal relationship has been difficult to measure. Social science research has swung backwards and forwards from `no effect' to `powerful effects' and back again. Violent video games are one of the latest research topics. `Computer games don't trigger aggression: report', said The West Australian, (December 4, 1999: 14). This five-year study, supported by the Australian government, concluded that `Computer games do not promote aggressive behaviour and are less disturbing for children than movies, according to a new report. They are not addictive, nor do they ruin family life, impair school performance or cause health problems' (Durkin and Aisbett, 1999). Cumberbatch and Howitt undertook a `state of the science' review of media effects in 1989. They covered research into specific effects in areas such as women and sex roles, race and racism, age and ageism, disable- ment, alcohol, prosocial issues, sex and violence. Their conclusion was that B A L N AV E S A N D C A P U T I 22

36. the `history of mass communication research is conspicuously lacking in any clear evidence on the precise influence of the mass media. Theories abound, examples multiply, but convincing facts that specific media content is reliably associated with particular effects have proved quite elusive' (Cumberbatch and Howitt, 1989: 25). One of the reasons for the difficulty in establishing cause and effect, according to Cumberbatch and Howitt, is that isolating the contribution mass media make to problems such as violence is extremely difficult because social perceptions about these phenomena are deeply embedded and often difficult to define. Lowrey and DeFleur (1983: 339) in their review of research argued that there are three general possibilities which describe a possible relationship between viewing media violence and aggressive behaviour: 1 The viewing of violence leads to aggressive tendencies. 2 Aggressive tendencies lead to the choosing of violence viewing. 3 Aggressive tendencies and the viewing of violence are both products of some third condition or set of conditions. Lowrey and DeFleur's conclusion is that `regular or frequent viewing of violent television programs may cause aggressive behaviour' (1983: 341). An analysis of other people's work, as the media and violence research demonstrates, requires a critical eye. Stern and Kalof (1996), summarized in Figure 2.1, provide a simple guide to checking `statements of fact' ± whether statements constitute genuine evidence. Some statements by authors are unsupported assertions. Simply saying that media violence leads to violent behaviour is an unsupported assertion. Simply referring to an author who says that media violence leads to violent behaviour is certainly trying to support the statement, but it is an `appeal to S TA R T I N G T H E I N Q U I R Y Unsupported assertion No Is anything presented as evidence? Appeal to Authority Yes No Casual observation Were any observations made about the facts? No Yes Were the observations reliable? Yes Evidence FIGURE 2.1 From assertions to evidence (used by permission) 23

37. authority' rather than substantial evidence. If the author referred to has substantiated his or her observations in a reliable manner, then we are getting closer to evidence. A literature review ± seeing what journal articles, books and other sources say about previous and contemporary research on the topic ± is essential for planning a research design. Like a compass, a literature review gives the researcher a bearing on what has been done. It also assists conceptualization and measurement in research design. For example, violence as we all know is real, but how does a researcher measure it? Sherman and Dominick (1986) attempted to measure the occur- rence of violence in music videos. In their content analysis of violence in music videos it was found that women, older adults and nonwhites are more likely to be represented as aggressors than victims. `Aggressors and victims were most likely to fall in the young adult age range (18±34). Interestingly, older adults (35±54) were nearly twice as likely to be aggressors as victims. For children, an opposite trend was noted: they were twice as likely to be the targets as the agents of aggression' (1986: 86±87). The unit of analysis in Sherman and Dominick's content analysis was the `concept video'. For measurement of violence they used Gerbner's definition of a violent act as an `overt expression of physical force with or without a weapon, against self or other, compelling action against one's will on pain of being hurt or killed or actually killing or hurting' (Gerbner, cited in Sherman and Dominick, 1986: 83±4). Independent coders then categor- ized aggressive and violent content by the type of violence (pushing, punching, slapping, gunshot, etc.), whether the aggressor was an individual or group, the sex, ethnic background and age of the victim, outcome of the aggression (death, injury, no effect) and whether the violence was portrayed realistically or unrealistically (1986: 84). Your reaction to Sherman and Dominick's definition of `violence' might be that it is too narrow ± surely there is emotional violence? psychological violence? what about shouting? Once you start raising these types of questions you are querying the validity of the definitions being used ± the construct validity and the internal validity of the study in question. These are exactly the types of questions you should raise about definitions in a study because it is these definitions that affect the measurement. Poor definitions mean poor measurement. The goal of a study should always be clearly stated. Spillman and Everington (1989), for example, provided uncomplicated sentences to describe their study: `The purpose of this study was to investigate whether the thin image presented by media today is consistent with earlier body- image stereotypes or whether current stereotypes are different. That is, is this image now perceived by college students as possessing more desirable characteristics?' Spillman and Everington (1989) found in their literature review that researchers have consistently reported stereotypical behavioural and personality traits associated with each of the three different body-build B A L N AV E S A N D C A P U T I 24

38. somatypes of endomorph, mesomorph and ectomorph. The most favourable traits have been assigned to the mesomorphic body build and the least favorable to the ectomorphic and endomorphic builds. Their sample was responses from 234 undergraduate students to a questionnaire. Each student was asked to fill out a questionnaire of 10 questions describing themselves (age, sex, weight range, height range, class, and school designa- tion). The next 24 questions required an individual to assign a behavioural characteristic to one of three female silhouettes. Spillman and Everington (1989) found in their results that the very thin body image, the ectomorph, was the most preferred somatype. Secondary Analysis Previous studies like Spillman and Everington's can be replicated ± done again ± in different contexts or at different points in time ± to see what is happening. Indeed, previous data can be re-analysed by other researchers. This is called secondary analysis. The wealth of survey data in modern societies, whether census, opinion poll, or other survey data, makes secondary analysis a common and economical method of research. Secondary analysis may also form part of exploratory work in a study ± for example identifying the characteristics of the people you are studying from previous survey data. If the same methodology for collecting data has been used over long periods of time, then useful comparison can be made. Time is, indeed, an important factor in social science research design. A thorough grounding in the literature ± the journal articles, books, reports ± associated with a research topic provides direction on how the research might be conducted. Knowing when to research is also a critical dimension of research design. KNOWING WHEN TO RESEARCH The Importance of Time Sherlock Holmes, quite rightly, calls his work `cases'. They are case studies that investigate `what is happening'. It is an investigation of one crime at a particular point of time. Holmes's only long-term criminal investigation involves his nemesis, Professor Moriarty ± the criminal genius. Exploring, describing and explaining in quantitative social science research requires a research design. Decisions on conceptualization, measurement and who is to be researched also require a research design. Time is important in a research design when the researcher wants to make compar- isons. A simple case study involves looking at a single phenomenon at a particular point in time. A longitudinal study involves investigating a phenomenon or different groups over time, to see if there are any relevant changes. For example, there have been changes in the demographics of the S TA R T I N G T H E I N Q U I R Y 25

39. journalism profession over time. Time is also relevant to experiments. An experiment is a very controlled attempt to see whether specific independent variables (assumed causes) are related to specific dependent variables (assumed effects). The researcher tries to manipulate the environment in an experiment to see what is happening. We will return to the idea of `independent variables' in Chapter 3. Table 2.2 gives an overview of the general options in research design. Exploratory and descriptive research often use cross-sectional designs. A British or US census, for example, is cross-sectional and describes the population at a given point of time. Explanatory research can use cross- sectional designs. A study designed to examine the influence of religious affiliation on voting, for example, could be done at a particular point in time. The problem with such `snapshot' research, if not replicated, is that it cannot take into account changes over time. Longitudinal research solves this problem by enabling comparison of the same groups and the same phenomena over time. TV ratings, for example, involve the establishment B A L N AV E S A N D C A P U T I TABLE 2.2 Time and research design Simple Case Study/Cross-Sectional Design What's happening? l A Longitudinal Study/Trends Studies/Panel Studies Has there been a change in A over time? l A l A Time1 Time 2 Comparison Are A and B different? l A l B Longitudinal Comparison Are A and B different over time? l A l A Time1 Time 2 l B l B Time1 Time 2 Longitudinal Comparison Are A and B different over time because of the independent variable? Experimental Group X (independent variable) öööö l A l A Time1 Time 2 Control Group l B l B Time1 Time 2 Adapted from Bouma (1993:111). (Used by permission). 26

40. of `panels', thousands of households, that are monitored regularly by ratings agencies. RESEARCH DESIGN We now have a rough checklist of the key elements of a research design. Great ideas are important in social science research, just as they are in detective fiction, but good ideas must be backed up by good design. Good design assists not only the investigation but the administration of quantitative research. It is very easy to underestimate the practical, admin- istrative, side of organizing, collecting and analysing evidence. Table 2.3 outlines the key elements associated with research planning. Many of the concepts in Table 2.3 have not yet been introduced, but they do provide you with a roadmap to what is required. We will return to all the major concepts as the book proceeds. SUMMARY Detectives try to make sense of what has happened, what is happening or what is going to happen, in the world about them. They use different methodologies and methods to make sense of, or add intelligibility to, the evidence that faces them. They are also required to show how they came to their conclusions and why one conclusion should be accepted rather than another. Social science research, similarly, is public. It is required to demon- strate its findings and to make itself open to critique and review. There is the possibility of error in social science research. These errors can be errors of human prejudice, errors of ego involvement (Holmes said he tried to put prejudice aside); errors in observation (Watson was chastised for not observing at all!); and premature closure of inquiry (Holmes was scathing about police prejudging crimes, Scarpetta never closed off investigations too quickly). There is also the possibility of overgeneralization (Father Brown never made wild conclusions from the evidence). The extracts from the works of Chesterton, Conan Doyle and Adams show that different processes of reasoning and explanation occur in detective fiction. These processes of reasoning have their parallel in quantitative research. Social scientists, like fictional detectives, have theories and hypotheses which they test against the evidence. They also have methods by which they collect and analyse that evidence. Theories, hypotheses and methods are integral to research and make research systematic. Of course, there are different theories and different methods in social science just as there are different styles of detection in fictional detective novels. But common to all detective fiction, and social science, is the requirement to explain what counts as evidence, and how conclusions fit S TA R T I N G T H E I N Q U I R Y 27

41. B A L N AV E S A N D C A P U T I TABLE 2.3 Organizing a quantitative research study PROBLEM What is the goal of the research? What is the problem, issue, or critical focus to be researched? What are the important terms? What do they mean? What is the significance of the problem? Doyou want to test a theory? Doyou want to extend a theory? Doyou want to test competing theories? Doyou want to test a method? Doyou want to replicate a previous study? Doyou want to correct previous research that was conducted in an inadequate manner? Doyou want to resolve inconsistent results from earlier studies? Doyou want to solve a practical problem? Doyou want to add to the body of knowledge in another manner? REVIEWOF LITERATURE What does previous research reveal about the problem? What is the theoretical framework for the investigation? Are there complementary or competing theoretical frameworks? What are the hypotheses and research questions that have emerged from the literature review? SAMPLE Who (what) will provide (constitute) the data for the research? What is the population being studied? Who will be the participants for the research? What sampling technique will be used? What materials and information are necessary to conduct the research? How will they be obtained? What special problems can be anticipated in acquiring needed materials and information? What are the limitations in the availability and reporting of materials and information? METHOD What methods or techniques will be used to collect the data? (This holds for applied and non-applied research) What procedures will be used to apply the methods or techniques? What are the limitations of these methods? What factors will affect the study's internal and external validity? Will anyethical principles be jeopardized? DATA ANALYSIS How will data be analysed? What statistics will be used? What criteria will be used to determine whether hypotheses are supported? What was discovered (about the goal, data, method, and data analysis) as a result of doing preliminary work (if conducted)? CONCLUDING INFORMATION How will the final research report be organized? (Outline) What sources have you examined thus far that pertain toyour study? (Reference list) What additional information does the reader need? What time frame (deadlines) have you established for collecting, analysing and presenting data? (Timetable) Adapted from Rubin et al. (1990) 28

42. the evidence. Logical argument based on sound empirical evidence for detectives and social scientists alike is superior to argument based on false evidence, or no evidence at all. Social scientists are detectives. They collect empirical evidence from the social world around them and make decisions about that evidence. `Empirical research' in this case simply means research based on evidence from the real world in contrast to theoretical, which refers to ideas that are abstract or purely analytical. `A theory remains theoretical until it is tested against the real world, with empirical evidence' (McNeill, 1985: 2). We will use the detective theme throughout this book on quantitative methods. Quantitative methods and statistics in social science are about logical inquiry. The different types of logic and the different types of personality in detective fiction and social science are often the same. You will find quirky characters in both types of detection, including the Dirk Gently. We will examine in depth the logic behind the styles of reasoning of detectives like Father Brown and Sherlock Holmes and the nature of the continuum between qualitative and quantitative. We will give examples of famous detective stories of social science ± showing how inquiries were started, investigated, and concluded. We will, finally, provide you with some important tools of analysis ± basic statistics. The Statistical Inquirer The Statistical Inquirer multimedia courseware provided on the CD-ROM introduces you to descriptive statistics. The step-by-step lessons cover the library, the concept of variables, measures of centrality, variance, correlation and regression, as well as the idea of statistical inference. These lessons are designed to reinforce what is covered in this text. Short videos, using LotusScreencam, provide a brief introduction to SPSS. A dataset from Patrick Rawstorne's doctoral study on subjective computer anxiety, Predicting and Explaining the use of Information Technology with Value Expectancy Models of Behaviour in Contexts of Mandatory Use, has been provided for practice. An Unsolved Mystery also gives you the opportunity to test your basic statistics skills. MAIN POINTS . There is a logic to social science inquiry. Quantitative social scientific inquiry requires evidence that is observable and testable. Researchers must give explanations for their conclusions. Social science research is public. This assists in avoiding bias in research. . A research design is the guide to how the research was constructed and carried out. S TA R T I N G T H E I N Q U I R Y 29

43. . Explanations in social science can be `nomothetic', dealing with the larger contexts in which social phenomena occur, or ìdiographic', dealing with unique events. `Data', however, is not `pregiven' in the social sciences ± it is a part of processes of interpretation in society. Definitions of `violence' are already affected, for example, by perceptions in society of what violence is. . Social science research can be exploratory, descriptive or explanatory. Exploratory research is used for testing new ideas or areas of research and/or testing methods used to collect data. Descriptive research, like TV ratings, describes `what is' and provides an overview of a phenomenon. Explanatory research seeks to explain `why'. `Why' questions tend to be theory questions. . Units of analysis are part of `what is counted' in quantitative research. This can be individuals or groups. It can also be items in a content analysis of a newspaper or a video. . A literature review is a preliminary investigation of what has been done or what is being done in an area of study. It is essential to planning a study. As you can see from the brief overviews of journalism as a profession and the relationship between media and effects, there is a substantial amount of existing research in these areas of study. . Secondary analysis is an economical way of finding out new things from existing data. . Time is a major factor in developing a research design. Research can be conducted once, at a particular point in time, or over time. Trend studies and panels are longitudinal and allow comparison over time. Experiments are the most controlled of all research studies, allowing investigation of the impact of specific phenomena over a period of time. REVIEW EXERCISES 1 Find two journal articles from a sociology, psychology or media journal that you think are examples of descriptive and explanatory research. Write a brief one-page summary of the research design. 2 Find a report on a research study in a newspaper, magazine or the internet that you think uses unsupported assertions, appeals to authority or casual observation (that do not appear to have been tested). 3 Classify the following statements as either unsupported assertion, appeal to authority, casual observation or evidence. Ì never met a person I did not like' Èveryone knows that smoking causes lung cancer' Àccording to Rogers (1970) opinion leaders have extraordinary influence over diffusion of innovations' B A L N AV E S A N D C A P U T I 30

44. REFERENCES Adams, D. (1987) Dirk Gently's Holistic Detective Agency. London: Heinemann. Babbie, E. (1986) The Practice of Social Research. Belmont: Wadsworth. Bouma, G.D. (1993) The Research Process. New York: Oxford. Boyd-Barrett, O. (1970) `Journalism recruitment and training: problems in professionalisation', in Jeremy Tunstall (ed.), Media Sociology. London: Constable. Chesterton, G.K. (1987) The Complete Father Brown: the enthralling adventures of fiction's best-loved amateur sleuth. London: Penguin. Cornwell, P. (1997) Cause of Death. London: Warner. Cumberbatch, G. and Howitt, D. (1989) A Measure of Uncertainty: The effects of the mass media. London: J. Libbey. Deutscher, I. (1973) What We Say, What We Do: sentiments and acts. Glenview, IL: Scott, Foresman. Doyle, Arthur Conan (1952) The Complete Sherlock Holmes. Garden City, New York: Doubleday. Durkin, K. and Aisbett, K. (1999) Computer Games and Australians Today. Office of Film and Literature Classification. Henningham, J. (1981) `The television journalist: A profile', Media Information Australia, 22: 5. Henningham, J. (ed.) (1990) Issues in Australian Journalism. Melbourne: Longman Cheshire. Henningham, J. (1993) `The hack's progress', Time, 11 January: 45. Henningham, J. (1996) `Australian journalists' views on professional associations', Asia Pacific Media Educator (1): 144±52. Hudson, W.J. (1964) `The education of a journalist in Australia', in E.L. French (ed.) Melbourne Studies in Education 1963. Melbourne: Melbourne University Press. pp. 321-47. Hudson, W.J. (1964) `Status of the metropolitan daily journalist in Australia', The Australian Journal of Social Issues, 2 (2). Johnstone, J.W.C, Slawski, E.J. and Bowman, W.W. (1976) The News People: A Sociological Portrait of American Journalists and Their Work. Urbana, IL: University of Illinois Press. Jones, R.L. and Swanson, C.E. (1954) `Small-city daily newspapermen: their abilities and interests', Journalism Quarterly, 31: 38±55. Judd, R.P. (1961) `The newspaper reporter in a suburban city', Journalism Quarterly, 38: 35-43. Lichter, S.R., Rothman, S. and Lichter, L.S. (1986) Media Elite: America's new power- brokers. Bethesda, MD: Ahler and Ahler. Lowrey, S. and DeFleur, M.L. (1983) Milestones in Mass Communication Research. New York: Longman. Martin, A. (1987) `Media and social change ± with special reference to television', Pacific Islands Communication Journal, 15 (1): 3±21. McNeill, P. (1985) Research Methods. London: Tavistock. Rubin, R.B., Rubin, A.M. and Piele, J. (1990) Communication Research: Strategies and Sources (2nd edn). Belmont, CA: Wadsworth. Sherman, B.L. and Dominick, J.R. (1986) `Violence and sex in music videos: TV and rock `n' roll', Journal of Communication, 36 (1): 79±93. S TA R T I N G T H E I N Q U I R Y 31

Introduction to Quantitative Research Methods_ An Investigative Approach.pdf

You are reading a preview.

Check these out next

Introduction to Quantitative Research Methods_ An Investigative Approach.pdf

More Related Content

Similar to Introduction to Quantitative Research Methods_ An Investigative Approach.pdf (20)

Recently uploaded (20)